Here is a link to what I think is a brilliant paper from some guys at MIT:
A biophysically-based neuromorphic model of spike rate- and timing-dependent plasticity
This is a very interesting paper in a number of ways.
Here is my synopsis of the paper:
The researchers made an analog model of the most basic mechanisms by which neurons communicate. They then used it to test a theory about the most fundamental aspects of neuronal interaction, and the role of endocannabinoids in this context. They found that endocannabinoids play a key role in modulating NMDA, the main exitatory neurotransmitter.
There has been an increasing interest in endocannabinoids among neuroscientists, as it becomes clear that they are not "minor" neurotransmitters. They are taking their rightful place among the "biggies" like NMDA, GABA, 5HT, dopamine, opioid, etc. Granted, they may exert a lot of their influence through modulation of the biggies, but this study shows that this modulation is of a co-equal nature.
This aspect of the research is interesting enough, for us herb lovers, but the implications are much bigger.
Usually when scientists model neuronal behavior, they do it with software. They use digital models that require massive computing power. This study is very different. They used a CMOS chip, not to digitally model the neurons, but to make an analog model.
This is a paradigm shift in neuronal modeling.
For you computer-savy folks, think of how much better hardware video or RAID is than software. This is a similar thing. The voltage channel potentials of the neurons are mimicked, not simulated, on the chip.
This has implications that are the stuff of science fiction.
This technology could be used in brain/machine interfaces.
Imagine a prosthetic eye that could not only restore sight to a blind person, but enhance vision to include magnification and zoom and the perception of wavelengths invisible to a "real" eye! Steve Austin cyborg stuff.
Or imagine controlling your computer, or your car, or any type of machine, the same way you control your hand.
On top of that, this technology could be used to make machine intelligence much more like ours than the current digital models.
I will stop now, if anyone is interested, I can elaborate more.
A biophysically-based neuromorphic model of spike rate- and timing-dependent plasticity
This is a very interesting paper in a number of ways.
Here is my synopsis of the paper:
The researchers made an analog model of the most basic mechanisms by which neurons communicate. They then used it to test a theory about the most fundamental aspects of neuronal interaction, and the role of endocannabinoids in this context. They found that endocannabinoids play a key role in modulating NMDA, the main exitatory neurotransmitter.
There has been an increasing interest in endocannabinoids among neuroscientists, as it becomes clear that they are not "minor" neurotransmitters. They are taking their rightful place among the "biggies" like NMDA, GABA, 5HT, dopamine, opioid, etc. Granted, they may exert a lot of their influence through modulation of the biggies, but this study shows that this modulation is of a co-equal nature.
This aspect of the research is interesting enough, for us herb lovers, but the implications are much bigger.
Usually when scientists model neuronal behavior, they do it with software. They use digital models that require massive computing power. This study is very different. They used a CMOS chip, not to digitally model the neurons, but to make an analog model.
This is a paradigm shift in neuronal modeling.
For you computer-savy folks, think of how much better hardware video or RAID is than software. This is a similar thing. The voltage channel potentials of the neurons are mimicked, not simulated, on the chip.
This has implications that are the stuff of science fiction.
This technology could be used in brain/machine interfaces.
Imagine a prosthetic eye that could not only restore sight to a blind person, but enhance vision to include magnification and zoom and the perception of wavelengths invisible to a "real" eye! Steve Austin cyborg stuff.
Or imagine controlling your computer, or your car, or any type of machine, the same way you control your hand.
On top of that, this technology could be used to make machine intelligence much more like ours than the current digital models.
I will stop now, if anyone is interested, I can elaborate more.
